See axolotl config

axolotl version: 0.4.1

base_model: mistralai/Mistral-7B-Instruct-v0.3
model_type: MistralForCausalLM
tokenizer_type: LlamaTokenizer

load_in_8bit: false
load_in_4bit: true
strict: false

datasets:
  - path: bshada/hw-infi
    type: alpaca
dataset_prepared_path: last_run_prepared
val_set_size: 0.1
output_dir: ./adapters/hw-infi-qlora-10-epoch

adapter: qlora
lora_model_dir:

sequence_len: 8192
sample_packing: false
pad_to_sequence_len: true

lora_r: 32
lora_alpha: 16
lora_dropout: 0.05
lora_target_linear: true
lora_fan_in_fan_out:
lora_target_modules:
  - gate_proj
  - down_proj
  - up_proj
  - q_proj
  - v_proj
  - k_proj
  - o_proj

wandb_project:
wandb_entity:
wandb_watch:
wandb_name:
wandb_log_model:

gradient_accumulation_steps: 1
micro_batch_size: 1
num_epochs: 10
optimizer: adamw_bnb_8bit
lr_scheduler: cosine
learning_rate: 0.0002

train_on_inputs: false
group_by_length: false
bf16: auto
fp16:
tf32: false

gradient_checkpointing: true
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true

loss_watchdog_threshold: 5.0
loss_watchdog_patience: 3

warmup_steps: 0
evals_per_epoch: 4
eval_table_size:
eval_max_new_tokens: 128
saves_per_epoch: 1
debug:
deepspeed:
weight_decay: 0.0
fsdp:
fsdp_config:
special_tokens:

adapters/hw-infi-qlora-10-epoch

This model is a fine-tuned version of mistralai/Mistral-7B-Instruct-v0.3 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.2173

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0002
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss
0.2156	0.5	1	0.2352
0.2098	1.0	2	0.2151
0.1621	1.5	3	0.1987
0.1491	2.0	4	0.1856
0.1256	2.5	5	0.1801
0.1092	3.0	6	0.1821
0.0901	3.5	7	0.1848
0.0996	4.0	8	0.1865
0.0633	4.5	9	0.1900
0.0804	5.0	10	0.1945
0.041	5.5	11	0.2015
0.063	6.0	12	0.2049
0.0526	6.5	13	0.2092
0.0302	7.0	14	0.2115
0.0398	7.5	15	0.2135
0.0191	8.0	16	0.2162
0.0334	8.5	17	0.2161
0.0157	9.0	18	0.2175
0.0307	9.5	19	0.2179
0.0143	10.0	20	0.2173

Framework versions

• PEFT 0.11.1
• Transformers 4.42.3
• Pytorch 2.3.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1